Dynamic neural activity during stress signals resilient coping

Abstract

Active coping underlies a healthy stress response, but neural processes supporting such resilient coping are not well-known. Using a brief, sustained exposure paradigm contrasting highly stressful, threatening, and violent stimuli versus nonaversive neutral visual stimuli in a functional magnetic resonance imaging (fMRI) study, we show significant subjective, physiologic, and endocrine increases and temporally related dynamically distinct patterns of neural activation in brain circuits underlying the stress response. First, stress-specific sustained increases in the amygdala, striatum, hypothalamus, midbrain, right insula, and right dorsolateral prefrontal cortex (DLPFC) regions supported the stress processing and reactivity circuit. Second, dynamic neural activation during stress versus neutral runs, showing early increases followed by later reduced activation in the ventrolateral prefrontal cortex (VLPFC), dorsal anterior cingulate cortex (dACC), left DLPFC, hippocampus, and left insula, suggested a stress adaptation response network. Finally, dynamic stress-specific mobilization of the ventromedial prefrontal cortex (VmPFC), marked by initial hypoactivity followed by increased VmPFC activation, pointed to the VmPFC as a key locus of the emotional and behavioral control network. Consistent with this finding, greater neural flexibility signals in the VmPFC during stress correlated with active coping ratings whereas lower dynamic activity in the VmPFC also predicted a higher level of maladaptive coping behaviors in real life, including binge alcohol intake, emotional eating, and frequency of arguments and fights. These findings demonstrate acute functional neuroplasticity during stress, with distinct and separable brain networks that underlie critical components of the stress response, and a specific role for VmPFC neuroflexibility in stress-resilient coping.

Keywords: binge alcohol intake; emotional eating; functional neuroimaging; resilience coping; stress.

Fig. 1.

Study design and subjective, physiologic, and neuroendocrine stress response. (A) A sample experimental successive run block made up of three baseline gray fixation runs, followed by six stress/neutral provocation runs and a 4-min recovery period to constitute each of the stress and neutral run blocks. (B) Significantly increased subjective stress and arousal ratings (visual analog 1–9 scale; P < 0.0001 each) and z-transformed scores for average heart rate (*P < 0.03) and plasma cortisol (*P = 0.05). ***P < 0.0001.

Fig. 2.

Correlation images from whole brain regression analysis showing association between stress-neutral (S-N) brain activity and cortisol response (S-N) are shown in A (P < 0.05, whole brain corrected) (see SI Appendix, Table S2 for MNI coordinates from the whole brain regression analyses). Corresponding correlation scatterplots from extracted beta weights of ROIs indicating areas of association from the S-N regression map and the S-N cortisol responses are shown in B. Red/yellow, positive correlation; blue-purple, negative correlation.

Fig. 3.

Significant condition × time period interactions in whole brain analysis showing time-dependent neural changes in brain response to stress. (A) Late–early runs of stress (first column), neutral (second column), and their contrast of stress-neutral (S-N) (third column) show increased dynamic activity for stress in the VmPFC, posterior cingulate cortex, and middle occipital gyrus, but decreased activity in the L ventrolateral PFC (VLPFC), insula, and superior/middle temporal gyrus, and no similar late-versus-early run changes in the neutral condition (second column). The S-N (third column) contrast indicates increased activity in the VmPFC and ventral striatum, midbrain, and L middle occipital gyrus, but decreased activity in the L VLPFC, insula, and superior/middle temporal gyrus for stress (late–early)–neutral (late–early) contrasts (P < 0.05, whole brain corrected). (B, i–iv) Time courses of responses across runs in key regions of interest (VmPFC, ventral striatum, L VLPFC, and insula) are shown to illustrate simple effects assessed in the whole brain contrasts of the interaction effects shown in A. Early, first two runs; Late, last two runs (of the successive six-run block for each condition); red/yellow, increased relative activation; blue/purple, decreased relative activation.

Fig. 4.

Whole brain functional connectivity with the VmPFC seed taken from the averaged brain stress response across all runs. Increased connectivity (shown in red/yellow) with the L anterior PFC (aPFC) and dorsolateral PFC (DLPFC) and L inferior parietal lobe (IPL) was found during stress average (S) compared with the neutral average (N) responses (P < 0.05, whole brain corrected).

Fig. 5.

VmPFC functional plasticity and coping. Individuals showing greater VmPFC activity in the late compared with early runs (A) during stress report higher active coping scores (B), and lower scores on emotional eating behaviors (C), lower nonbinge levels of alcoholic drinks consumed per occasion (D), and low to never getting into arguments and fights with others (E) (see SI Appendix, Table S1 and Detailed Materials and Methods for measures used).